Genetic Slippage in Response to Sex

-It is widely held that sexual reproduction enhances the potential for evolution by expanding the range of variation on which selection can act. However, when nonadditive gene action contributes to the expression of traits under selection, sexual reproduction can also result in slippage of the mean genotypic value in the direction contrary to selection. We show how the magnitude of genetic slippage depends on the extent to which segregation and recombination create maladapted genotypes. We also show how random mating can induce a change in the expressed genetic variance for a quantitative trait by eliminating Hardy-Weinberg disequilibria and reducing gametic-phase disequilibria. Depending on whether genes of like effects are in repulsion or coupling disequilibrium, this change will be positive or negative. Thus, depending on the mode of gene action and the form of the selection function, sexual reproduction can either enhance or impede the short-term response of quantitative characters to selection. Although this issue is relevant to all sexual populations, it is most easily investigated in species that infrequently engage in sex, since prolonged phases of clonal propagation can greatly exaggerate genetic disequilibria. We describe a population of the cyclical parthenogen Daphnia pulex in which sexual reproduction induced average changes in the means of life-history characters equivalent to approximately one-tenth of a phenotypic standard deviation. Contrary to the usual expectation, sex also caused a significant reduction in the expressed genetic variance for several traits in this population. A large fraction of the genetic variance in Daphnia appears to be due to dominance, and in the study population, clonal selection appears to cause a buildup of coupling disequilibrium between genes and gene combinations of like effects. Species that reproduce both sexually and asexually provide several advantages for investigating the genetic architecture of quantitative characters. First, through clonal propagation, it is possible to estimate the total expressed genetic variance for any quantitative trait by partitioning the phenotypic variance into its withinand between-clone components. Under a suitable experimental design, the former contains all of the environmental variance for the trait, and the latter, all of the genetic variance. Second, the change in the genetic variance between a parental generation and that of its sexually produced progeny provides information on the amount of quantitative variation that was hidden (or exaggerated) in the parental generation by Hardy-Weinberg and gametic-phase disequilibria (Lynch and Gabriel 1983; Lynch 1984). Third, something close to the expressed additive genetic variance can be estimated from the covariance between parents and their sexually produced progeny. By combining these three types of analysis, insight can be gained on the extent to which genetic disequilibria and nonadditive sources of * E-mail: [email protected]. Am. Nat. 1994. Vol. 144, pp. 242-261 I O 1994 by The University of Ch~cago. rights reserved 243 GENETIC SLIPPAGE IN RESPONSE TO SEX